A substantial portion of primary care patients being treated with short-acting opioids for chronic non- cancer pain also develop opioid dependence, and their addiction often worsens with continued use of these drugs. Non-opioid medications frequently are inadequate for pain control in these patients, and physicians have limited options for referring them to substance use treatment. Many physicians, therefore, have tried a middle-ground approach of switching these patients to the long-acting opioids methadone and buprenorphine, which have analgesic effects but lower potential for exacerbating addiction. Experience with this has been encouraging, but the evidence is largely anecdotal and research is needed to help guide clinicians in making evidence-based treatment decisions for this growing patient population. The purpose of this pilot randomized trial is to compare methadone and buprenorphine for pain control and other clinical outcomes in patients with chronic pain and coexisting addiction. Primary care physicians at 6 community practices in Buffalo, NY, will refer 66 patients who have objective evidence to support the diagnosis of a chronic non-cancer pain condition related to the spine following a surgical procedure (i.e., failed back syndrome) and who also have an addiction to prescription opioids. Consenting participants will be assigned to one of two treatment conditions: 1) methadone 30-60 mg per day in divided doses (active comparator) or 2) sublingual buprenorphine 4-16 mg per day in divided doses (experimental group). All participants will also be referred for behavioral therapy. The patients will receive the medications for 6 months at an outpatient center under the supervision of a board-certified Addiction Medicine physician. During that time they will keep a treatment diary in which they use established assessment measures to record daily pain levels, and they will have monthly face-to-face meetings with study personnel to gather additional information, including use of other drugs, quality of life, attendance at behavioral counseling, overall level of function, and urine toxicology. Patients will then return to their original primary care providers for ongoing pain management, under arrangements with the study Addiction Medicine physician. Final data will be collected at the nine-month point to assess pain and other outcomes, including whether patients have remained in treatment. Study findings will provide preliminary but important clinical guidance on the use of buprenorphine versus methadone for patients with chronic pain and coexistent addiction. Buprenorphine has a good safety profile, so if it is found to be at least as clinically effective as methadone, this would constitute a valuable contribution to evidence-based treatment decisions. The findings also will contribute to subsequent research supporting development of clinical guidelines. The ultimate impact would be a paradigm shift from limited specialty options to mainstream medical management so patients with chronic pain and coexistent addiction could be treated in a manner similar to those with other chronic conditions.